Almost every sentences uses the passive voice. I think the use of "can,"
"are often," "do not have to be," and so on, is an awkward attempt to
work around the use of the keywords as currently defined, or as redefined.
It's not, actually; partly, it's to keep the tone similar to the
current RFC 2119, but partly, well, I think it reads well this way. I
could propose the following version that recasts the sentence into the
active voice. Does that help you?:
=== NEW ===
Many standards track documents use several words to signify
the requirements in the specification. Authors often capitalize
these words, as shown below, but they do not have to. This
document defines how these words are interpreted in IETF documents
when the words are capitalized.
o Authors can use these words as defined here, but they are not
required to. Specifically, normative text does not require the
use of these key words. Authors may use them for clarity and
consistency when they want that, but a lot of normative text does
not use them, and does not need to use them.
o The words have the meanings specified herein only when they are
capitalized.
o When these words are not capitalized, they have their normal
English meanings; this document has nothing to do with them.
Authors who follow these guidelines should incorporate this phrase
near the beginning of their document:
Readers are to interpret the key words "MUST", "MUST NOT",
"SHOULD", "SHOULD NOT", and "MAY" in this document as
BCP 14 [RFC2119],[RFCxxxx] describes them when, and only
when, they appear capitalized, as shown.
=== END ===
How does this update affect interpretation of existing documents?
It does not; that's why you need to use both references, to show that
both 2119 and this document apply. Documents that only cite 2119 only
get 2119 applied.
It's ironic, I think, that the active voice clause, "Authors who follow these
guidelines should incorporate this phrase" uses an uncapitalized "should,"
and I think it should be a "MUST."
There's no irony here: this document intentionally does not use BCP 14
key words, neither does 2119, and I think they should not. I'd rather
that BCP 14 not be self-referential.
To Scott's semi-serious suggestion to deprecate "SHOULD," I'm hesitant.
If there were another word for "do this unless you have a really good
reason not to," I'd go for it. "Reasons not to" must be documented
wherever possible, though sometimes not all reasons can be foreseen
(see what I did there?).
Used passive voice, you mean? And it reads just fine, really.
Barry